Employment Oriented Macroeconomic Policy and the Challenge of Outsourcing

Gerald Epstein

Professor of Economics and Co-Director, Political Economy Research Institute

(PERI)

University of Massachusetts, Amherst

February, 2006

Paper prepared for the Conference on “The New Global Division of Labor: Winners and Losers from Offshore Outsourcing”, Center for Global Initiatives, Mount Holyoke College, March 3-4, 2006. Significant parts of this paper are drawn from joint work with James Burke of Mt. Holyoke College. I thank him for his important contributions. Minsik Choi of the University of Massachusetts, Boston also made important contributions. I also thank Arjun Jayadev, Derek Weener and Elissa Braunstein for help. The Political Economy Research Institute provided research support. All errors are mine, of course.

I. Introduction

There is no longer much doubt that globalization and global integration are having

significant impacts on the trajectories of many economies around the world. But there is

still little agreement about what that impact is. Several years ago, I and my co-authors

identified a number of possible trajectories found in the debate, and most of these

possibilities still seem relevant today (Crotty, Epstein, Kelly, 1998).1

The first is "The Race to the Bottom". (Bluestone and Harrison, 1982; Greider,

1997). According to this view, capital will increasingly be able to play workers,

communities and nations off against one another as they demand tax, regulation and wage

concessions while threatening to move. According to this view, increased mobility of

multinational corporations (MNC's) benefits capital while workers and communities lose.

A modified version is that the winners in the race to the bottom will include highly

educated (or skilled) workers, or workers in particular MNC rent appropriating

professions (e.g. lawyers and investment bankers), along with the capitalists; the losers

will be unskilled workers and the unemployed.

The second view, "The Climb to the Top", is the opposite of the first. It suggests

that multinational corporations are attracted less by low wages and taxes than by highly

educated workers, good infrastructure, high levels of demand and agglomeration effects

arising from the existence of other companies who have already located in a particular

place. According to this view, competition among states for foreign direct investment

(FDI) will lead counties in both the North and the South to try to provide well educated

labor and high quality infrastructure in order to retain and attract foreign

investment.(Reich, 1992; Bhagwati, 2004; Friedman, 2004) Thus footloose capital and

competition, far from creating a race to the bottom will induce a climb to the top around

the world.

This climb to the top could lead to the outcome represented by the third view:

"Neo-liberal Convergence". This is the widely held mainstream claim that free mobility

of multinational corporations, in the context of de-regulation and free trade, will produce

increased living standards in all countries. This process will, moreover, transfer capital 1 See Crotty, Epstein and Kelly (1998) for much more discussion of these trajectories and the argument, more generally, in this first section.

1

and technology abroad, thereby raising the standards of living of those in the poorer

countries at a faster rate than those in the wealthier ones, eventually generating a world

wide convergence in living standards. This may result from the process of competition

for capital described above, or simply from the market processes of dissemination of

capital and technology throughout the globe.

These same processes could, however lead to the outcome envisaged in the fourth

view, "Uneven Development". "Uneven Development" has a long and, now, ironic

history: it holds that one region of the world will grow at the expense of another region.

Of course, for decades, the dominant version of this view was the theory of imperialism:

if the South integrated itself with the North, the North would grow at the expense of the

South. Now, the fear seems to be the opposite: by having to compete with cheap Southern

labor, an integrated world economy will help the South grow, but this time at the expense

of the North.

Outsourcing/off-shoring in manufacturing and now increasingly in services reflect

newer production and sourcing strategies by major companies, many of them

multinational corporations. Hence, these phenomena raise many of the same issues

discussed above. In particular, does outsourcing/offshoring make one or the other of these

trajectories more likely?2

As one might expect, there are advocates for each of the positions. As Alan

Blinder reminds us in a recent paper, “One thing you should never predict is the future”

(Blinder, 2005). Nonetheless, in the area of off-shoring and outsourcing,one cannot avoid

rejecting this sage advice since offshoring, particularly in services, is more a thing of the

future, then it is of the present or past.

The major point I wish to make here is the same point we made with respect to

FDI more generally (Crotty, Epstein and Kelly, 1998). Offshoring is not inherently good

or bad. It depends on the context. In particular, which of the four views best approximates

reality will strongly depend on the overall national and international context within which

offshoring occurs. In particular, I focus here on three aspects of the overall context which

are especially important in determining the impact of offshoring: the state of aggregate

2In line with emerging practice, I will mostly use the term “offshoring” or “foreign outsourcing” to refer to outsourcing from abroad.

2

demand (AD), the nature of the domestic and international rules of the game and

institutions governing investment, and the nature of domestic and international

competition. These three factors have a significant impact on the effects of offshoring on

the economy, and in particular on their effects on wages, inequality and the level of

unemployment.

More concretely, when offshoring occurs in a context of high levels of aggregate

demand and effective rules of the game which in turn limit the destructive aspects of

competition, then it may indeed have a positive impact on nations and communities. On

the other hand, when it occurs in a context of low levels of aggregate demand and

destructive economic and political competition in the absence of effective rules of the

game, then it can have a significantly negative impact on workers in both home and host

countries.

While the mechanisms through which these three factors condition the impacts of

offshoring are myriad, I focus on two: the effect of these three factors on the bargaining

power of firms relative to workers, nations and communities; and their contribution to

coordination problems that hinder the ability of governments at all levels to make policies

which can capture the benefits from offshoring. The increased relative bargaining power

of capital means that their leaving or even the threat to leave can lead to reductions in

wages, worsening of working conditions for workers and low tax rates and revenues for

governments. As for coordination problems, this weakened bargaining position and these

weakened rules of the game make it more difficult for communities and nations to avoid

"Prisoner's Dilemma" solutions where countries' wages, tax rates, expenditure choices

and regulatory structures are severely distorted from the point of view of the community

as a whole as they try to compete for capital and contracts. Indeed, they may become so

distorted that they are suboptimal from the point of view of the corporations themselves.

As a result, in this context, off-shoring can contribute in a significant way to the problems

of unemployment, wage stagnation and inequality.

Currently, the dominant governance framework in the global economy can be

defined as “neo-liberal” (eg. Baker, Epstein, Pollin, 1998; Stiglitz, 2002) where

macroeconomic austerity, privatization, financial liberalization, and trade liberalization

are promoted by international officials, business people, governments and many

3

economists. Like some others, I believe that in the current "Neo-liberal" regime, makes a

race to the bottom much more likely than many mainstream economists believe. This

view stems from the claim that within the Neo-liberal regime, there are strong forces

which lead to insufficient levels of aggregate demand and therefore chronic

unemployment, coercive competition, and destructive domestic and international rules of

the game -- that is, precisely those factors which undermine the potentially positive

effects of offshoring.

Among the most important of these factors is the government revenue and

governance implications of capital mobility generally, and competition for off-shoring in

particular. A number of authors, believe that as offshoring expands, it may lead to large

scale disruptions in developed country labor markets, disruptions that might call for

significant government interventions, including broadening the social safety net (eg.

Blinder, 2005; Jensen and Kletzer, 2005; Madrick and Milberg, 2006 ). Yet, the very

processes of globalization and off-shoring themselves might undermine both the capacity

and the will of governments to intervene in these ways. In this case, the race to the

bottom scenario becomes more likely.

To redress this balance, policies and activities must be undertaken to enhance the

bargaining power of states and citizens as globalization and offshoring expands, in order

for states to be able to have the ability and “willingness” to compensate losers. These

policies might even need to include forms of “protection” or at least, states must retain

the ability to “threaten” to protect in order to preserve the power to compensate losers. In

this context – where the full political economy of globalization and power is taken into

account – some forms of protection may be found to be efficient.

The rest of the paper is organized as follows. In the next section, I present some

new findings, developed in conjunction with James Burke, on the scope and impacts of

outsourcing in the manufacturing sector in the U.S. These results show that

manufacturing outsourcing in the US has been accelerating in recent years, is associated

with job losses in a number of U.S. manufacturing industries, and that it does seem to be

associated with unit labor cost differentials between the U.S. and off-shoring host

countries. In section III, I briefly discuss service sector off-shoring, briefly summarizing

some key information from the work of others. In the final section, I discuss the impacts

4

of these processes on the capacity and willingness of the government to undertake

necessary policies to reduce the costs and spread the benefits of the changes.

3II. Foreign Outsourcing and Intermediate Goods Trade In Manufacturing

There has been much more research on the dimensions and impacts of off-shoring

from the manufacturing sector in developed countries than on service sector off-shoring

both because the former has been occurring for a longer period of time and also the data

to study the phenomenon are much better. Building on the work of Feenstra and Hansen

and Campa and Goldberg, most of this empirical research has looked at manufacturing

off-shoring as leading to a rise in the international trade of intermediate goods. As inputs

produced globally take the place of intermediate stages of production at home, we

witness an increased flow of imports of intermediate goods across country borders. Past

studies have calculated the importation of intermediate goods relative to total

intermediate input purchases as an indicator of foreign outsourcing activity in an industry

sector (see, for example, Feenstra and Hanson, 1999 and Campa and Goldberg 1997).

While changes in the share of imports in total intermediate goods will not fully capture

the extent of globalization of production in an industry – some foreign outsourcing

activity by US firms will show itself as a displacement of US production of final goods or

exports rather than an increase in imports of intermediate goods – it does provide a

measurable indicator that can be tied directly to important channels offshoring activity.

Tracking the share of imported intermediate manufactured goods in total purchases of

intermediate manufactured goods allows us to discern changes in a significant part of

foreign outsourcing in the US manufacturing sector over time.4

3 This section draws liberally on my joint paper with James Burke, “Rising Foreign Outsourcing and Employment Losses in US Manufacturing, 1987 - 2003", PERI, 2006. 4 There are other possible sources of increasing imported inputs in US production in addition to outsourcing activities by US firms. First, if foreign firms set up production in US sites, they are likely to use intermediate goods shipped from their home countries or other foreign suppliers. These activities would increase the shares of imported inputs in US production without any new outsourcing activity by US firms. Second, a rise in the relative price of domestic versus foreign inputs can lead to a rise in the value share of imported inputs without actually representing a shift in the location of production abroad.

5

Feenstra and Hanson (1999) find that imported intermediate goods have increased

from 5.3% of total intermediate purchases for U.S. manufacturing industries in 1972 to

7.3 % in 1979, and 12.1% in 1990. Using a narrower measure of intermediate goods,

Campa and Goldberg (1997) provide evidence for Canada, Japan, the United Kingdom

and the United States in the mid-1970s, mid-1980s and mid-1990s. They find that

imported inputs have increased from 4.1% of total intermediate goods in 1975 to 6.2 % in

1985, and 8.2% in 1995 for U.S. manufacturing industries.

In this paper, we will present our own measure of imports of intermediate goods

in manufacturing industries and include a more recent period, covering years from 1987

to 2003. We then look at the relationship between our measure of foreign outsourcing

activity and job loss and wages in US manufacturing industries in recent years. We also

explore how foreign outsourcing activity in US industries is related to the unit labor costs

in foreign industries.

Measuring Imported Inputs in US Manufacturing Industries

We use data provided in the Bureau of Economic Analysis’s national input-output

accounts to calculate our measures of imported intermediate goods. We measure

imported intermediate goods used by US manufacturing industries using a similar

methodology as the Feenstra and Hanson the Campos and Goldberg studies mentioned

above. That is, we begin by finding the import share of each commodity (the share of the

commodity used in the US economy that is imported) as well as the value of each

commodity used in the production process of each industry. For each industry, we then

multiply the value of the commodity used in production by the import share of that

commodity to find the value of imported inputs of the commodity used by that industry.5

In future work, we will test the size of some of these effects but we assume for now that these effects are small compared to the effect of outsourcing activity. 5 A basic assumption of this method of calculating the value of imported inputs is that the import share of the commodity when it is used as an intermediate good in each particular industry is the same as the import share of the commodity in the economy as a whole as calculated from the I/O accounts.

6

By summing up the imported inputs of each commodity used by that industry, we can

find the industry’s total imported inputs used in production. The industry data we require

to carry out these calculations are included in the ‘use tables’ of the BEA’s input-output

accounts. These tables show how industries use inputs of commodities to produce goods

in the economy and also report the quantities of commodities that are imported into the

US.6

Imports of intermediate goods used in manufacturing industry production

We calculate the share of imported goods in total purchases of intermediate

manufactured goods for manufacturing groups and for the manufacturing sector as a

whole for the years 1987, 1992, 1997, and for 1998 through 2003. We consider only

intermediate goods that are manufactured commodities.7

Chart 1 shows the share of imported inputs in total inputs of manufactured goods

used in production for nineteen manufacturing industry groups and for the manufacturing

sector as a whole in 1987 and 2003. For every industry group and for the manufacturing

sector as a whole, the share of imported inputs used in production has risen substantially

over the time period. For all manufacturing, the share of imported inputs rose from 12.4

percent to 22.7 percent between 1987 and 2003. The industry groups with the highest

measures of foreign outsourcing activity in the use of inputs were the Apparel/Leather

Products group, the Computer/Electronic Products group, and the Motor Vehicles/Bodies

and Trailers/Parts group. In these three industry groups, imported inputs made up about

one-third of all manufactured inputs used in production in 2003. 6 In 1997, the Bureau of Economic Analysis began to provide industry input-output tables using the North American Industry Classification System (NAICS); in the years previous to 1997, the input-output accounts used the Standard Industrial Classification (SIC) system. The data we present and use in our analysis breaks the manufacturing sector into the nineteen manufacturing groups defined by the NAICS; for years previous to 1997, we need to allocate the industry groups defined by the SIC system into these nineteen NAICS groups. Although the transition from the SIC system to the NAICS does not allow a perfect allocation of group data across industrial classifications, we strove to minimize the distortion of industry data as much as possible. Refer to our paper to see how industry data was converted from the SIC system into NAICS defined groups. 7 For more detailed information, see Burke and Epstein (2006).

7

Table 1 shows the share of imported inputs in total inputs used in production for

the whole manufacturing sector and manufacturing industry groups in 1987, 1992, 1997,

2002, and 2003. Between 1987 and 2003, the industry groups with the largest increases

in the share of imported inputs used in production were the Apparel/Leather Products

group (15.4%), the Textiles group (15.2%), the Motor Vehicles/Bodies and Trailers/Parts

group (13.9%), the Plastics and Rubber Products group (11.9%), and the

Computer/Electronic Products group (11.4%). Table 1 also shows that the growth in the

share of imported inputs in the manufacturing sector as a whole accelerated in the later

part of the 1987 to 2003 period. Of the total increase of 10.4 percentage points in the

import share of inputs used in production in the sector as a whole, the earliest period

(1987 - 1992) accounts for 1.5 percentage points, the middle 5-year period (1992 - 1997)

accounts for 3.8 percentage points, and the latest 6-year period (1997 – 2003) accounts

for 5.0 percentage points. Faster growth in the share of imported inputs in the most

recent 1997 – 2003 period is also seen in 13 of the 19 manufacturing industry groups.

The increase in foreign outsourcing in the latest period was especially fast for the Motor

Vehicles/Bodies and Trailers/Parts industry group in which the years from 1997 – 2003

accounted for over three-quarters of the increase in the share of foreign-sourced inputs.

Of the 13.9 percentage point increase in that group’s imported input share between 1987

and 2003, from 16.3 percent to 30.2 percent, the most recent 6-year period accounted for

11.1 percentage points.

8

Chart 1: Imported inputs of manufactured goods used in production,as share of total inputs used in industry, 1987-2003.

0%

5%

10%

15%

20%

25%

30%

35%

40%

All

Man

ufac

turin

g

Com

pute

r and

ele

ctro

nic

prod

ucts

App

arel

and

leat

her a

nd a

llied

prod

ucts

Mot

or v

ehic

les,

bod

ies

and

trai

lers

, and

par

tsM

isce

llane

ous

man

ufac

turin

gEl

ectr

ical

equ

ipm

ent,

appl

ianc

es, a

nd c

ompo

nent

sO

ther

tran

spor

tatio

n eq

uipm

ent

Text

ile m

ills

and

text

ile p

rodu

ctm

ills

Mac

hine

ry

Prim

ary

met

als

Che

mic

al p

rodu

cts

Plas

tics

and

rubb

er p

rodu

cts

Fabr

icat

ed m

etal

pro

duct

s

Woo

d pr

oduc

ts

Non

met

allic

min

eral

pro

duct

s

Furn

iture

and

rela

ted

prod

ucts

Prin

ting

and

rela

ted

supp

ort

activ

ities

Pape

r pro

duct

s

Petr

oleu

m a

nd c

oal p

rodu

cts

Food

and

bev

erag

e an

dto

bacc

o pr

oduc

ts

1987 2003

9

10

Table 1: Imported inputs of manufactured goods used in production in US manufacturing industries,

as a share of total inputs used by industry, various years.

Industry Imported Inputs as a Share of Total Inputs

1987 1992 1997 2002 2003 Change in Share,

1987 - 2003

All Manufacturing 12.4% 13.9% 17.7%

22.3%

22.7% 10.4%

Apparel and leather and allied products 18.9% 24.1% 24.5% 32.3% 34.2% 15.4%

Computer and electronic products 22.3% 26.5% 32.7% 34.6% 33.7% 11.4%

Motor vehicles, bodies and trailers, and

parts 16.3% 18.0% 19.1% 28.7% 30.2% 13.9%

Textile mills and textile product mills 8.8% 11.3% 14.3% 22.5% 24.0% 15.2%

Miscellaneous manufacturing 16.5% 18.6% 18.0% 23.6% 23.8% 7.3%

Electrical equipment, appliances, and

components 13.2% 14.8% 18.3% 23.1% 23.5% 10.3%

Machinery 13.0% 13.9% 17.1% 22.1% 23.1% 10.1%

Other transportation equipment 12.6% 15.9% 18.5% 23.5% 22.3% 9.7%

Primary metals 12.8% 14.3% 21.2% 21.3% 21.3% 8.5%

Chemical products 10.6% 12.0% 15.7% 20.5% 21.1% 10.5%

Plastics and rubber products 9.0% 10.6% 13.3% 20.2% 21.0% 11.9%

Fabricated metal products 12.4% 12.5% 15.8% 18.8% 19.2% 6.8%

Nonmetallic mineral products 9.9% 10.4% 13.8% 17.4% 18.3% 8.3%

Wood products 8.9% 8.8% 14.3% 17.8% 17.9% 9.0%

Furniture and related products 10.1% 10.6% 13.1% 17.1% 17.7% 7.6%

Printing and related support activities 9.4% 8.1% 14.9% 15.3% 16.0% 6.7%

Paper products 10.6% 10.3% 15.2% 15.1% 15.3% 4.7%

Petroleum and coal products 9.5% 8.5% 9.4% 12.8% 13.2% 3.7%

Food and beverage and tobacco products 5.8% 6.1% 6.5% 9.8% 10.5% 4.7%

Imports of intermediate goods produced in manufacturing industries

In addition to examining how industries use imported intermediate goods in

production, we also measure the degree to which imported inputs compete with the

production of intermediate goods by US manufacturing industry groups. That is, we look

at the phenomenon of foreign outsourcing from the perspective of the industry making

intermediate goods as well as the perspective of the industry using intermediate goods in

production. We do this because firms often use intermediate goods that are not produced

in their own industry group. Consequently, foreign outsourcing that involves shifting

purchases of intermediate goods from domestic to foreign suppliers can raise the share of

imported inputs used in production while not directly displacing production in a firm’s

own industry group. Instead, the demand for production (and workers) will fall in other

industries as a result of this kind of foreign outsourcing activity. Because we are

interested in data that will allow us to explore links between foreign outsourcing and

industry employment at home, we want to identify the industries where these

manufactured inputs are produced as well as where they are used.

Using the “make tables” of the BEA’s industry input-output accounts, we find the

share of production of each manufactured commodity attributable to each of the nineteen

major manufacturing industry groups. 8 First, we refer to the calculations described

above, which drew on the use tables, to find the amounts of each commodity that are used

as inputs to production and the share of these inputs that are imported. We then assign

these commodity values according to each industry’s production or ‘make’ share.

Summing across all commodities for the imported intermediate goods assigned to an

industry, we find the total value of imports among the intermediate goods produced by

8 For example, all commodities categorized as ‘Apparel, leather and allied products’ are produced by the following manufacturing industry groups in these shares: Food and beverage and tobacco products 0.51% Textile mills and textile product mills 0.95% Apparel and leather and allied products 96.80% Printing and related support activities 0.81% Electrical equipment, appliances, and components 0.12%

11

the industry. We find these industry group calculations for 1998 through 2003, the years

for which the required data was available in a consistent way from the BEA. We are

especially in this time period, as manufacturing sector employment fell precipitously in

the US during this time (from 17.5 million workers in 1998 to 14.5 million workers five

years later). 9 Note that in our discussion of employment and unit labor costs below, we

use these “make tables”.

Chart 2 shows the import share of the intermediate manufactured goods produced

by US manufacturing groups for the manufacturing sector as a whole and for nineteen

manufacturing industry groups in 1998 and 2003. For every industry group and for the

manufacturing sector as a whole, the share of imports in total inputs produced has risen

over the time period. For all manufacturing, the share of imported inputs rose from 19.1

percent to 22.8 percent between 1998 and 2003.10 The industry group facing the highest

share of imports in the supply of manufactured inputs it produces was the

Apparel/Leather Products group – almost two-thirds of inputs produced by this group is

produced in foreign sites. Other industry groups producing manufactured inputs with

high import shares were the Computer/Electronic Products group, the Miscellaneous

Manufacturing group, the Motor Vehicles/Bodies and Trailers/Parts group, and the

Electrical Equipment, Appliances and Components group. In these four industry groups,

imported inputs made up one-third or more of the supply of the manufactured inputs they

produced in 2003.

Table 2 shows the import share of total inputs produced by the whole

manufacturing sector and manufacturing industry groups for the years between 1998 and

2003. For these years, the industry groups with the largest increases in the share of

9 Current Employment Statistics, US Bureau of Labor Statistics. 10 For the manufacturing sector as a whole in 2003, the import share for inputs produced by industries in Chart 2 is essentially the same as the one shown in Chart 1 for inputs used in production (22.8% and 22.7%, respectively) . This makes sense because almost all manufactured inputs used in the manufacturing sector as a whole are also produced in the manufacturing sector. The differences between imported input shares in Charts 1 and 2 (and Tables 1 and 2) show up when comparing the manufacturing industry groups. As discussed in the text, this is because industries often use inputs in their production which are produced in another industry.

12

imported inputs used in production were the Furniture and Related Products group, the

Apparel/Leather Products group, the Electrical Equipment, Appliances and Components

group, the Motor Vehicles/Bodies and Trailers/Parts group, and the Textile Mills and

Textile Product Mills group. For these groups, the import share of the total supply of

manufactured inputs produced rose by between eight and ten percentage points in five

years.

Foreign outsourcing and shifts in demand and supply within US manufacturing

The rising share of imported inputs in US manufacturing has taken place in the

context of shifts in both the level and composition of domestic demand and supply in the

manufacturing sector. Chart 3 shows the levels of domestic and foreign-sourced

intermediate manufacturing goods used in US manufacturing production from 1987 to

2003 in constant 2003 dollars. In the earlier period, from 1987 to 1997, the share of

imported intermediate goods rose as their use in production grew at a faster rate than the

growth in the use of domestic inputs. Between 1987 and 1997, foreign-sourced inputs

13

Chart 2: Imported inputs of manufactured goods used in US manufacturing production, as a share of total inputs produced by industry, 1998 and 2003

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

All

Man

ufac

turin

g

App

arel

and

leat

her a

nd a

llied

pro

duct

s

Com

pute

r and

ele

ctro

nic

prod

ucts

Mis

cella

neou

s m

anuf

actu

ring

Mot

or v

ehic

les,

bod

ies

and

trai

lers

, and

par

ts

Elec

tric

al e

quip

men

t, ap

plia

nces

, and

com

pone

nts

Mac

hine

ry

Furn

iture

and

rela

ted

prod

ucts

Che

mic

al p

rodu

cts

Text

ile m

ills

and

text

ile p

rodu

ct m

ills

Prim

ary

met

als

Oth

er tr

ansp

orta

tion

equi

pmen

t

Woo

d pr

oduc

ts

Non

met

allic

min

eral

pro

duct

s

Petr

oleu

m a

nd c

oal p

rodu

cts

Fabr

icat

ed m

etal

pro

duct

s

Plas

tics

and

rubb

er p

rodu

cts

Pape

r pro

duct

s

Food

and

bev

erag

e an

d to

bacc

o pr

oduc

ts

Prin

ting

and

rela

ted

supp

ort a

ctiv

ities

19982003

14

15

Table 2: Imported inputs of manufactured goods used in production in US manufacturing industries,

as share of total inputs made by industry, various years.

Industry Imported inputs as a share of total inputs

1998 1999 2000 2001 2002 2003

Change in Share,

1988-2003

All Manufacturing 19.1% 20.0% 21.8% 21.9% 22.3% 22.8% 3.6%

Apparel and leather and allied products 54.3% 56.2% 59.7% 63.2% 62.5% 64.0% 9.7%

Computer and electronic products 36.3% 37.5% 40.2% 39.2% 41.0% 39.4% 3.1%

Miscellaneous manufacturing 32.1% 33.3% 35.1% 35.7% 36.2% 35.5% 3.3%

Motor vehicles, bodies and trailers, and

parts 26.7% 27.8% 31.3% 32.9% 33.7% 35.4% 8.7%

Electrical equipment, appliances, and

components 25.2% 27.3% 29.6% 31.2% 33.0% 34.7% 9.6%

Machinery 25.5% 26.1% 27.5% 27.6% 26.9% 29.4% 3.8%

Furniture and related products 15.1% 17.3% 19.6% 20.0% 22.5% 25.0% 9.9%

Chemical products 17.3% 18.2% 19.7% 21.1% 22.7% 23.4% 6.1%

Textile mills and textile product mills 15.0% 16.1% 18.0% 19.4% 21.1% 23.0% 8.0%

Primary metals 20.7% 20.1% 23.0% 22.5% 21.5% 21.3% 0.6%

Other transportation equipment 19.6% 20.9% 24.2% 24.1% 21.1% 19.9% 0.3%

Wood products 14.1% 15.5% 15.4% 16.2% 16.7% 16.6% 2.5%

Nonmetallic mineral products 12.5% 13.5% 14.4% 14.0% 15.3% 16.5% 3.9%

Petroleum and coal products 9.2% 9.5% 11.4% 12.2% 12.1% 13.0% 3.8%

Fabricated metal products 9.5% 10.1% 10.9% 11.1% 11.8% 12.5% 3.0%

Plastics and rubber products 10.2% 10.6% 11.0% 11.2% 11.5% 12.2% 2.0%

Paper products 10.9% 11.3% 12.1% 12.2% 12.2% 12.1% 1.1%

Food and beverage and tobacco

products 5.9% 6.2% 6.4% 6.6% 7.0% 7.7% 1.8%

Printing and related support activities 2.3% 2.7% 2.8% 2.7% 3.1% 3.5% 1.2%

grew by 75 percent in real terms (from $140 billion to $245 billion) while domestic

inputs grew by just 12 percent (from $988 to $1,117 billion). Imported intermediate

goods grew more slowly in real terms in the later period from 1998 to 2003, but their

share in total inputs continued to grow as the volume of domestic inputs fell. During

these years of poor growth in US manufacturing production, the real value of total inputs

used in US manufacturing fell by 9 percent. In this later period, the value of foreign-

sourced intermediate goods rose by 15 percent (from $245 billion to $282 billion) while

the value of domestic inputs fell by 14 percent (from $1,117 billion to $957 billion).

Table 3 focuses on the recent period from 1998 to 2003 and allows us to look

more closely at the demand and supply of both final and intermediate manufacturing

goods. Real production by US manufacturing industries grew by just 3.3 percent during

these five years. While domestic demand for all manufactured goods rose by 8.5 percent

in real terms in the period, three-quarters of the growth in demand was met by the rise in

imported goods. The supply of all imported manufactured goods rose by 32 percent

while the supply of all domestically produced goods rose by 3 percent during this time.

For final goods, domestic demand increased by 17 percent during this period, of which

almost half was met by the growth of imported final goods (domestic supply rose by 11

percent while imports of final goods rose by 40 percent).

The domestic demand for intermediate goods for production by the US

manufacturing sector fell by 9 percent during these years. Given the 3.3 percent rise in

manufacturing output, this 9 percent drop in demand for inputs of manufactured goods

represents the effect of technological change and rising productivity in manufacturing

production in relation to inputs of manufactured goods. While the domestic demand for

manufactured inputs fell, foreign sourced inputs continued to grow. Thus, of the $159

billion dollar drop in the sale of domestically produced intermediate goods between 1998

and 2003, we can attribute about three-quarters to the $122 billion fall in domestic

demand for inputs due to rising productivity and about one-quarter to the $37 billion

dollar growth in the foreign outsourcing of inputs.

16

Chart 3: Domestic and Imported Inputs to US Manufacturing Industries, 1987-2003

2003 dollars

0

200

400

600

800

1,000

1,200

1,400

1987 1992 1998 2003

$ B

illio

ns

Domestic Inputs

Imported Inputs

1987 1992 1998 2003

Dollars 988 1,014 1,117 957 Domestic

Inputs Share 87.6% 86.1% 82.0% 77.2%

Dollars 140

164

245

282

Imported

Inputs

Share 12.4% 13.9% 18.0% 22.8%

Total Inputs 1,128 1,178 1,362 1,239

17

Table 3: Change in domestic demand, domestic supply, imports, and exports

for US manufacturing industries, 1998-2003.

(billions of 2003 dollars)

Percent

change,

1998-

2003

Change,

Foreign Outsourcing and Employment Loss in US Manufacturing

1998 2003 1998-

2003

Total domestic demand

(shipments – exports + imports) 4,127.3 4,476.6 349.3 8.5%

Domestic supply 3,326.5 3,419.2 92.7 2.8%

Imports 800.8 1,057.4 256.6 32.0%

Domestic demand for final goods 2765.4 3,236.8 471.4 17.0%

Domestic supply 2,209.9 2,461.8 251.9 11.4%

Imports 555.5 775.0 219.5 39.5%

Domestic demand for

intermediate goods used in

manufacturing production

1,361.9 1,239.8 -122.2 -9.0%

Domestic supply 1,116.6 957.4 -159.2 -14.3%

Imports 245.3 282.4 37.1 15.1%

Export demand 543.2 577.0 33.7 6.2%

Total US production (exports

and domestic supply) 3,869.7 3,996.2 126.5 3.3%

18

Employment in US manufacturing has fallen precipitously since the late 1990s.

Following a more gradual downward trend since the late 1970s, between 1998 and 2003

employment in manufacturing fell by 18 percent, declining by over 3 million jobs in five

years to reach its lowest level in over half a century (see Chart 4). There is considerable

debate among economists and policy makers about the causes of this dramatic dislocation

of manufacturing workers in recent years.11 Along with the growth of foreign

outsourcing, mounting import competition is also put forward as playing a substantial

role in the large-scale downsizing of US manufacturing employment. Both of these

potential pressures on US manufacturing are often associated with the rapid growth of

industrial production in developing counties in the last decade.12 Rising labor

productivity and structural changes in the economy leading to shifts in the composition of

demand away from purchases of manufactured goods are also pointed to as factors

drawing down US manufacturing employment. In this section, we use our estimated

measures of imported intermediate goods to explore how the level of foreign outsourcing

activity and other factors are related to the drop in employment in US manufacturing

industries in the years from 1998 to 2003.13

11 See, for example, these articles for a range of views on the sharp decline in manufacturing employment in recent years: Martin Baily and Robert Lawrence, 2004; and Josh Bivens, 2004 12 Among the developing countries that have attracted the most attention as new competitors for US producers are China and India. For these countries, industrial production grew by 16 and 12 percent annually over the last year (Emerging Market Indicators, The Economist, September10, 2005). Imports from China to the US grew by about 30 percent between 2003 and 2004. 13 Two estimates of the size of outsourcing activity and its effects on employment have received attention recently. First, reports produced by the Forrester Research consulting firm in 2002 and 2004 have projected job loss in service industries over the next decade as the result of outsourcing. However, the lead author of the Forrester Research reports has described them in press interviews as based on “a very rough and gross calculation” and “educated guesses”. A second recent estimate of the effects of outsourcing on US jobs has come from the addition of a new question added this year to the Bureau of Labor Statistics’ (BLS) Mass Layoffs Survey. This question asked firms carrying out layoffs of more than fifty workers whether these layoffs were the result of relocating production to foreign sites. The estimate of job loss related to outsourcing derived from the BLS’s Mass Layoffs Survey, however, has significant limitations. Most notably, the Survey

19

Table 4 shows data on employment together with the share of imports in the

intermediate goods produced by nineteen industry groups that make up the manufacturing

sector for the period from 1998 to 2003. Employment fell in all nineteen industry groups

during these years, although the rates of employment decline varied greatly across

industries. The Apparel and Leather and Allied Products group, which saw its workforce

cut in half during this period, experienced the largest drop in employment; the industry

group with the smallest decline in employment was the Food, Beverage and Tobacco

Products group with a fall of just 3.5 percent.

We begin a preliminary analysis of the relationship between outsourcing activity

and job loss by examining the correlation between the import share of intermediate goods

produced in each of the nineteen industry groups and the change in industry employment

between 1998 and 2003. Table 5 presents both Pearson and rank-order correlation

results for 1998 to 2003 between employment change and outsourcing activity. We

correlate employment change with both 1) the level of the import share of inputs in 2003,

and, 2) the 1998 – 2003 change in the import share of inputs. The results show a highly

significant and moderate correlation (ρ = -0.656) between the import share of inputs in

2003 and the decline in industry employment between 1998 and 2003. Industries with

high levels of outsourcing activity in 2003 tended to experience higher levels of

employment loss since 1998. This result is not apparent from simply ranking industries

by outsourcing levels and employment losses - the rank-order correlation between the

import share of inputs in 2003 and employment decline is not significant. The 1998 to

only covers a small fraction of all job losses each quarter, namely job losses from large layoffs taking place over a limited span of time. In addition, this survey doesn’t capture outsourcing that involves replacing the firm’s domestic production with imports of intermediate goods or services, or outsourcing that replaces the firm’s domestic suppliers of intermediate goods or services with foreign suppliers. Finally, we question whether the employers included in the BLS survey fully report how many of their laid-off workers are replaced by workers in their foreign production sites.

20

Chart 4: US Manufacturing Employment, 1987-2003

10000

11000

12000

13000

14000

15000

16000

17000

18000

19000

1987 1989 1991 1993 1995 1997 1999 2001 2003

Thou

sand

sds

of W

orke

rs

2003 change in the import share of inputs is not significantly correlated with industry

employment decline, measured either with the Pearson or the rank-order correlation.

These results suggest that the industries with high employment losses from 1998 to 2003

tend to have high levels of outsourcing activity, but high employment losses in an

industry over the period are not associated with growth in the extent of outsourcing

activity.

We continue our analysis by comparing the extent of foreign outsourcing activity

in 2003 for the six industries with the highest rate of job loss between 1998 and 2003 to a

group made up of the remaining thirteen industries. Testing supported using the division

of the sample into 6 high job loss industries versus 13 low job loss industries. A t-test

procedure indicated that the average rate of job loss in the top six job loss industries was

21

significantly different from the average for the remaining 13 groups, with t = -5.231,

significant at the 0.01 level.14 Additionally, of all possible partitions of the sample, this

14 The independent samples t-test is used to test whether the means of two groups are significantly different. Testing supported using the division of the sample into 6 high job loss industries versus 13 low job loss industries - of all possible partitions of the sample, this one yielded a t-test score with the highest significance level.

22

Table 4: Employment and imported intermediate goods as a share of inputs produced, 1998-2003 (employment in thousands)

Industry Name Employment 1998

Employment2003

Employment Change,

Employment Change, %,

98-03

Imports Share of Inputs,

Imports Share of Inputs Change,

98-03 2003 98-03

Apparel and leather

and allied products 682.2 339.3 -342.9 -50.3% 64.0% 9.7%

Textile mills and

textile product mills 629.3 417.6 -211.7 -33.6% 23.0% 8.0%

Computer and

electronic products 1,793.4 1,319.1 -474.3 -26.4% 39.4% 3.1%

Primary metals 632.6 466.1 -166.5 -26.3% 21.3% 0.6%

Machinery 1,488.7 1,131.6 -357.1 -24.0% 29.4% 3.8%

Electrical equipment,

appliances, and

components

590.3 450.1 -140.2 -23.8% 34.7% 9.6%

Other transportation

equipment 805.0 644.5 -160.5 -19.9% 19.9% 0.3%

Paper products 619.1 503.3 -115.8 -18.7% 12.1% 1.1%

Printing and related

support activities 823.1 671.5 -151.6 -18.4% 3.5% 1.2%

Petroleum and coal

products 132.7 111.4 -21.3 -16.1% 13.0% 3.8%

Fabricated metal

products 1,733.3 1472 -261.3 -15.1% 12.5% 3.0%

Plastics and rubber

products 940.5 805.4 -135.1 -14.4% 12.2% 2.0%

Motor vehicles,

bodies and trailers,

and parts

1,301.7 1,124.5 -177.2 -13.6% 35.4% 8.7%

Furniture and related

products 651.4 569.6 -81.8 -12.6% 25.0% 9.9%

Wood products 614.9 540.3 -74.6 -12.1% 16.6% 2.5%

Miscellaneous

manufacturing 727.4 653.9 -73.5 -10.1% 35.5% 3.3%

Chemical products 989.5 891.7 -97.8 -9.9% 23.4% 6.1%

Nonmetallic mineral

products 539.8 492 -47.8 -8.9% 16.5% 3.9%

23 Food and beverage

and tobacco

products

1,764.2 1,702.9 -61.3 -3.5% 7.7% 1.8%

Table 5: Pearson and Rank-Order Correlation of Employment Change and Import Share

of Intermediate Goods, 1998 - 2003

Percent Employment Change 1998-

2003 and…. Correlation coefficient Significance

Pearson

Correlation -0.656*** .002

Import Share of

Inputs in 2003 Rank-Order

Correlation 0.344 .149

Pearson

Correlation -0.343 .151

Change in Import

Share of Inputs,

1998 - 2003 Rank-Order

Correlation 0.021 .932

Observations: 19

*Significant at 10% level; ** Significant at 5% level; *** Significant at 1% level

Rank-order correlations show the Spearman’s rho correlation statistic.

Import share in total manufacturing inputs produced by industry.

one yielded a t-test score with the highest significance level. The six industries in the

highest job loss group (with their 1998 – 2003 employment declines in parentheses) are:

Apparel/Leather Products (50.3 percent), Textiles 33.6 percent), Computer/Electronic

Products (26.4 percent), Primary Metals (26.3 percent), Machinery (24.0 percent), and

Electrical equipment/ Appliances/ Components (23.8 percent). These six high job loss

industry groups made up 33 percent of total manufacturing sector employment in 1998

and accounted for 54 percent of the total decline of over 3.1 million jobs in

manufacturing employment between 1998 and 2003.

Chart 5 shows the average for the import share of manufactured inputs produced

in 2003 for the high and low job loss groups. The average of the import share of inputs

for the six high job loss industry groups was 35.3 percent versus 17.9 percent for the low

24

job loss industry groups. A t-test procedure indicated that the average share of imported

inputs for the high job loss group is significantly different from the average for the low

job loss group with t = 2.989, significant at the 0.01 level. This result is consistent with

the view that high rates of job loss are related to higher levels of foreign outsourcing of

production activity in US industries.

We now look at how growth in the level of import shares of intermediate goods is

related to employment loss for manufacturing industries. Chart 6 shows the average for

the percentage point change in the import share of manufactured inputs from 1998 to

2003 in the high and low job loss groups. For 1998 to 2003, the average change in the

import share of inputs was 5.8 percent for the six high job loss industry groups versus 3.7

percent for the low job loss industry groups. A t-test procedure indicated that there is not

a significant difference in these percentage changes between the high job loss group and

the low job loss group (t = 1.357, significant at the 0.20 level). These results suggest that,

while high rates of job loss are related to high levels of foreign outsourcing across

industries between 1998 and 2003, high growth rates in foreign outsourcing are not

related to high job loss during these years. This is consistent with our earlier analyses of

correlation between foreign outsourcing and employment loss discussed above.

25

Chart 5: Manufacturing job loss in 1998-2003 and foreign outsourcing

Average share of imports in total manufacturing inputs produced by industry in 2003

17.9%

35.3%

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%Imported Inputs

Share

Low Job Loss IndustriesHigh Job Loss Industries

The six industries in the ‘high job loss’ group are: Computer/Electronic Products; Apparel/Leather Products; Machinery; Textiles; Primary Metals; and Electrical Equipment/Appliances/Components. The remaining thirteen industries make up the ‘low job loss’ group.

26

Chart 6: Manufacturing job loss in 1998-2003 and foreign outsourcing

Average growth in share of imports in total manufacturing inputs produced by industry in 2003

3.7%

5.8%

0.0%

2.0%

4.0%

6.0%

8.0%

10.0%Growth in

Imported Inputs Share

Low Job Loss IndustriesHigh Job Loss Industries

The six industries in the ‘high job loss’ group are: Computer/Electronic Products; Apparel/Leather Products; Machinery; Textiles; Primary Metals; and Electrical Equipment/Appliances/Components. The remaining thirteen industries make up the ‘low job loss’ group.

Regression Results

Burke and Epstein (2006) report econometric results that support the basic result outlined

here. We find that technology, export demand, and outsourcing all play a role in the

decline in manufacturing employment. While overall, increased import competition

accounts for a relatively small share of the decline in employment (about 5% of the

decline) in some industries, the contribution is more economically significant in selected

industries, including textiles and motor vehicles. Importantly, domestic and foreign

export demand play a larger role than does outsourcing, per se. This suggests that

measures to expand aggregate demand at home and abroad can play a useful role in

ameliorating the negative impacts of offshoring in manufacturing.

Foreign Outsourcing and Labor Costs

27

A major question surrounding foreign outsourcing in US manufacturing is

whether employers shift production outside US borders in response to lower labor costs

abroad, especially in developing countries. In this section, we turn our focus towards

how the level of foreign outsourcing activity by US firms is associated with differences in

labor costs in US production versus production abroad. More specifically, we explore

the relationship between our measure of an industry’s foreign outsourcing - the share of

imported intermediate goods in total inputs - and the ratio of foreign to US unit labor

costs in that industry.

The unit labor cost is defined as worker compensation per hour divided by value-

added per hour.15 The unit labor cost can be interpreted as the labor cost of producing a

dollar’s worth of value-added in production. We use data provided in the Industrial

Statistics Database compiled by the United Nations Industrial Development Organization

(UNIDO) to calculate our measures of US and foreign unit labor costs. The UNIDO

Industrial Statistics Database provides industry data for 180 countries on annual wages

paid and annual value-added. We divide annual wages by annual value-added to

calculate the industry’s unit labor cost for that country in that year. For each industry, we

select a group of countries that appear most likely to host new production sites for

intermediate goods used in US production. The criteria we use to select these countries

are that they 1) account for more than a 2 percent share of US imports for the industry,

and 2) have shown positive growth in US import share for the industry over the 1998 to

2003 period. To calculate the foreign unit labor cost for an industry, we sum the unit

labor costs of the countries in the selected group, after weighting each country’s unit

labor cost by their share of total imports in the group.

Although UNIDO strives to minimize cross-country incompatibilities in the data,

some issues still remain that are important to our measures of unit labor cost. First, since

wages are reported on a yearly basis rather than hourly, some distortion arises because

there is variation across countries in how many hours are worked per year. Next, the data

15 Equivalently, the unit labor cost can be described as the hourly wage divided by hourly labor productivity, when productivity is measured in relation to value-added.

28

generally only reflect wages and usually do not include other labor costs in total

compensation, such as health or retirement benefits. These non-wage compensation costs

vary across countries.

The most current UNIDO Industrial Statistics Database does not report wages for

all of the years in our study period (1998 through 2003). For most of the countries

utilized, the latest year for which we are able to calculate a unit labor cost is 2000 or

2001; for a handful of countries within certain industries, we can only calculate the unit

labor cost from the mid to late 1990s. As a proxy for the unit labor costs in later years in

the period, we use the unit labor cost for the last year available. Also, because the

UNIDO database does not include wage data for China, we gather this information from

the China Statistical Yearbook to calculate that country’s unit labor cost. While the

UNIDO data has these limitations, it remains the best data source for our purposes in

terms of its country and industry coverage and the variables provided. 16

We begin a preliminary analysis of the relationship between outsourcing activity

and unit labor costs in foreign sites by examining the correlation between the import

share of total intermediate goods produced in each of the nineteen industry groups and

the foreign/US ratio of unit labor costs in 1998 in each industry. Table 6 presents both

Pearson and rank-order correlation results for outsourcing activity and the foreign/US

ratio of unit labor costs. We correlate the ratio of unit labor costs in 1998 with both 1)

the import share of total inputs produced in 2003, and, 2) the 1998 – 2003 change in the

import share of inputs. Of the four correlations shown, only the rank correlation between

the foreign/US ratio of unit labor costs and the 1998 – 2003 change in the import share of

inputs shows a significant result. In this case, we find a significant (at the 10% level) and

moderate correlation between the change in the import share of inputs between 1998 and

2003 and the measure of relative unit labor costs in foreign versus US locations (ρ =

0.407). This result offers some evidence that industries facing lower levels of unit labor

costs in foreign locations tended to experience higher growth in outsourcing activity

16 The ILO’s International Labour Statistics database also provides information on industry wages and value-added, but the coverage is more limited. The US Bureau of Labor Statistics’ Foreign Labor Statistics database provides unit labor cost data, but coverage is limited to a small set of richer countries.

29

between 1998 and 2003. Table 6 does not provide evidence that lower unit labor costs

abroad are associated with higher levels of outsourcing activity in 2003. According to

both bivariate and rank correlation, there is not a significant relationship between the

industry level of imported inputs in total inputs in 2003 and the ratio of foreign to

domestic unit labor costs for the industry.

We continue our analysis by comparing the average level of outsourcing activity

in 2003 between 1) a ‘high foreign labor cost’ group made up of the seven industries in

which weighted unit labor cost in foreign locations is equal or greater than the US unit

labor cost in 1998, and 2) a ‘low foreign labor cost’ group made up of the remaining

twelve industries with weighted unit labor costs in foreign locations lower than the US

unit labor cost. For the

Table 6: Pearson and Rank-Order Correlation of Foreign/US Unit Labor Cost Ratio in

1998 and Import Share of Intermediate Goods, 1998 – 2003

Ratio of Foreign to US Unit Labor

Costs in Industry Groups in 1998 Correlation coefficient Significance

and ….

Pearson

Correlation -0.328 .171

Import Share of

Inputs in 2003 Rank-Order

Correlation 0.305 .204

Pearson

Correlation -0.280 .246

Change in Import

Share of Inputs,

1998 - 2003 Rank-Order

Correlation 0.407* .084

Observations: 19

*Significant at 10% level; ** Significant at 5% level

Rank-order correlations show the Spearman’s rho correlation statistic.

Import share in total manufacturing inputs produced by industry.

30

industries in the ‘high foreign labor cost’ group, the weighted unit labor cost in foreign

locations in 1998 ranges from 99 to 167 percent of the US unit labor cost. In the ‘low

foreign labor cost’ group, the weighted foreign unit labor cost is 43 to 90 percent of the

US level in 1998. A t-test procedure indicates that the average level of the foreign unit

labor cost in 1998 in the seven ‘high foreign labor cost’ industries is significantly

different from the average for the remaining 12 industries, with t = -5.104, significant at

the 0.01 level.17 The seven industries in the ‘high foreign labor cost’ group are: the

Other Transportation Equipment group, the Paper Products group, the Primary Metals

Manufacturing group, the Printing and Related Support Activities group, the Food and

Beverage and Tobacco Products group, the Petroleum and Coal Products group, and the

Chemical Products group. These seven industry groups made up 33 percent of total

manufacturing sector employment in 1998.

Chart 7 shows the average for the import share of manufactured inputs produced

in 2003 for the high and low foreign labor cost groups. The average of the import share

of inputs for the seven industry groups with foreign unit labor cost equal or greater than

the US level was 14.4 percent versus 28.7 percent for the industry groups with unit labor

costs lower than the US level. A t-test procedure indicated that the average share of

imported inputs for the high foreign labor cost group is significantly different from the

average for the low foreign labor cost group with t = 2.384, significant at the 0.05 level.

This result is consistent with the view that lower labor costs in foreign production sites

are associated with higher levels of foreign outsourcing of production activity in US

manufacturing industry groups.

Next we look at how growth in outsourcing activity is related to the relative cost

of labor in foreign locations for manufacturing industries. Chart 8 shows the average for

the percentage point change in the import share of manufactured inputs from 1998 to

2003 in the high and low foreign labor cost groups. For 1998 to 2003, the average

change in the import share of inputs was 2.1 percent for the seven high foreign labor cost

17 The independent samples t-test is used to test whether the means of two groups are significantly different.

31

industry groups versus 5.6 percent for the low foreign labor cost industry groups. A t-test

procedure indicated that there is a significant difference in these percentage changes

between the high foreign labor cost group and the low foreign labor cost group (t = 2.889,

significant at the 0.01 level). These results suggest that high growth rates in foreign

outsourcing activity between 1998 and 2003 for industry groups are associated with low

labor costs in foreign production sites relative to production at home.

Chart 7: Ratio of Foreign to US Unit Labor Cost in 1998 and Foreign Outsourcing in 2003

Average share of inputs in total manufacturing inputs produced by industry.

28.7%

14.4%

0.0%

5.0%

10.0%

15.0%

20.0%

25.0%

30.0%

35.0%

High Foreign Labor Cost Low Foreign Labor Cost

Imported Inputs Share

For the ‘high foreign labor cost’ group, the foreign/US unit labor cost ratio ranges from 0.99 to 1.67. For the ‘low foreign labor cost’ group, the foreign/US unit labor cost ratio is between 0.43 and 0.90.

The seven industries in the ‘high foreign labor cost’ group are: Other Transportation Equipment; Paper Products; Primary Metals Manufacturing; Printing and Related Support Activities; Food and Beverage and Tobacco Products; Petroleum and Coal Products; and Chemical Products. The remaining twelve industries make up the ‘low foreign labor cost’ group.

32

Chart 8: Ratio of Foreign to US Unit Labor Cost in 1998 and Foreign Outsourcing in 2003

Average growth in the share of inputs in total manufacturing inputs produced by industry.

2.1%

5.6%

0.0%

2.0%

4.0%

6.0%

8.0%

10.0%

High Foreign Labor Cost Low Foreign Labor Cost

Growth in Imported Inputs Share

For the ‘high foreign labor cost’ group, the foreign/US unit labor

cost ratio ranges from 0.99 to 1.67. For the ‘low foreign labor cost’ group, the foreign/US unit labor cost ratio is between 0.43 and 0.90.

The seven industries in the ‘high foreign labor cost’ group are: Other Transportation Equipment; Paper Products; Primary Metals Manufacturing; Printing and Related Support Activities; Food and Beverage and Tobacco Products; Petroleum and Coal Products; and Chemical Products. The remaining twelve industries make up the ‘low foreign labor cost’ group.

Regression Analysis on Foreign Outsourcing and Labor Costs Burke and Epstein (2006) present regression analysis of the impacts of relative labor

costs on foreign outsourcing. Again, the results support the bi-variate results discussed so

far. U.S. firms do respond to relative unit labor costs when making outsourcing decisions.

33

Such responses, all else equal, are consistent with pressures that contribute to a “race to

the bottom”.

Conclusions on Impacts of Outsourcing in Manufacturing

Our empirical suggests that manufacturing outsourcing has increased significantly in

recent years and that, moreover, it has contributed to employment losses, especially in

industries where foreign outsourcing has been concentrated. Firm choices to engage in

outsourcing does seem to respond to relative unit labor costs at home and abroad, which

is consistent with the idea that relative wage competition could put pressure on wages,

and/or employment at home. Still, aggregate demand measures, at home and abroad, are

even more significant in determining employment losses in manufacturing, and

technology changes are also important. The important role of aggregate demand has

important policy implications, which I draw out in somewhat more detail below.

III. Off-Shoring in the Services Sector

A newer phenomenon of increasing concern, but also of less clear dimensions, is the

spread of off-shoring to the service sector. Other papers in this conference have focused

on this issue so I will not attempt to cover the same in detail. Here I will make only a few

remarks.

While there is enormous uncertainty about how widespread service off-shoring

will become, several authors have made compelling arguments that it is likely to become

much more widespread over the next ten to twenty years than it is today. Through careful

empirical analysis, Jensen and Kletzer (2005) find that a significant number of service

industries and occupations that traditionally appear to be un-tradable, are likely to

become tradable in the new world of digital commerce. Moreover, workers in these

tradable services generally have higher skill levels and are paid more than workers in

tradable manufacturing or in non-tradable service sectors. In addition, they find that

workers in tradable services that are displaced are different from manufacturing workers

who are displaced by trade, in that, they too have higher skill and earnings than those

displaced from manufacturing. In short, domestic employment loss is evidently occurring

34

as a result of services offshoring as well as manufacturing offshoring, as we described

above. Moreover, workers with higher skills and wages are evidently being affected in

the case of service sector offshoring.

Alan Blinder (2005) speculates that this phenomenon could become much more

widespread and significant in the future. He argues that in the future, “impersonal

services” will be increasingly tradable, placing higher paid US, European and Japanese

workers in direct competition with much lower paid workers in developing countries,

especially those who are able to communicate in Western languages. These pressures will

increasingly put jobs and wages at risk in the richer countries and will require significant

labor market adjustments and call for significant interventions by governments,

especially in the area of education and social safety nets.

Blinder and others argue strongly against “protectionism” in the sense of policies

that will interfere with trade in these goods and services. They argue that interfering with

trade will lower global, if not national, welfare, and that, moreover, in the cases of

electronic offshoring, such protectionism is likely to be impossible.

Some economists from off-shoring hosts such as India doubt that this offshoring

is likely to have large and widely dispersed benefits in their home countries

(Chandrasekhar and Gosh, 2006). Their argument hinges partly on the claim that large

multinational firms will increasingly enter these activities and capture the lion share of

the value added from them, leaving the host country workers and capitalists with a

relatively small share of the benefits.

Milberg, et. al. (2005), also raise the key question of profit accumulation from

off-shoring activities. They argue that the key determinant of the impact of off-shoring on

the overall economy will stem, not from static efficiency gains from increases in the

division of labor, but from the dynamic effects (or lack thereof) of employment

generating impacts of investment undertaken by firms making profit gains from these

activities.

I believe this issue, in fact, is crucial, but one can broaden the frame. As Milberg,

et. al, argue, it will make a great deal of difference how much and where MNC’s invest

these increased profits, and how much and what type of employment is generated by

these investments. But, it will also make a crucial difference as to whether governments

35

can capture tax revenues from these profits to bolster social safety nets and social

investments, and also, whether the political economy of such valuable activities will give

governments the ability to implement tax and regulatory policies to manage off-shoring

to the benefit of more than a small minority of share-holders of MNC’s. However, with

MNC’s having large increases in profits at stake, they will have political capital available

and the incentive to use it to ward off costly regulations and tax bills. These

considerations, therefore, take us firmly into the realm of the political economy of

redistributive and regulatory policy as a necessary part of our discussion of offshoring.

IV. Impact of Globalization and Off-Shoring on the Ability and Willingness of the

State to Spread the Benefits and Reduce the Costs

As discussed above, off-shoring appears to be associated with increasing profit

rates and shares for capital (Milberg, et. al., 2005). More generally, in the US and many

other OECD countries, in the last decade labor shares have declined significantly and in

many cases this decline has been associated with a rise in profit shares.

These declines support the view that neo-liberal globalization, accompanied as it

is by capital mobility, and now, its cousin, off-shoring, is shifting the bargaining power

toward the owners of mobile firms, and away from immobile labor and states. (eg,

Rodrik, 1997, 1998; Burke and Epstein, 2002; among many others). The same processes

are making it more difficult to tax capital income, and may be shifting the incidence of

taxation from capital to labor in many OECD countries (See Epstein, 2000, for a survey

of this literature; and Altshuler and Grubert, 2005, for more recent work.) As Altshuler

and Grubert point out, at least three parties play a role in this “race to the bottom”: host

governments, home governments and multinational companies. (See also Muti, 2003, for

a good survey.)

These problems are made more severe by the ways in which the “perceived”

ability of MNC’s to shift production abroad, either by FDI or by off-shoring, enables

companies to issue threats to governments and workers. These threats, in turn, make it

costly for workers, unions and governments to implement institutional changes that

reduce or threaten to reduce companies’ profits or prerogatives. These threat effects have

36

been recognized to some extent in the literature (Bronfenbrenner, 20000; Burke and

Epstein, 2000 and Choi, 2004, for example.).

These points are highly relevant to the issue of the impact of offshoring. Many

economists have argued for strengthening the social safety net in developed countries as

off-shoring and other types of production shifting accelerates. In other words, Rodrik

notes, when globalization increases, there is a perceived increase in the demand for social

protection (Rodrik, 1997). At the same, our political economy argument suggests that

increased globalization might also reduce the ability and the willingness of the state to

provide these social protections, because globalization reduces tax revenues available to

the state and reduces the power and appetite of the state to impose restrictions or

regulations on capital. By this reasoning, increased globalization reduces the supply of

social protection. Figure 1 illustrates this dilemma (See Braunstein and Epstein, 2001,

from which this diagram was taken, for more discussion of these relationships). It shows

that increased openness or globalization increases the “demand” for social protection

while, at the same time, reducing the “supply” of social protection.

37

Figure 1Demand for and Supply of Social Protection

SocialProtection

Openness

G0G1

Demand

Supply

}StruggleSP0

Demand: workers and citizens from firms and the stateSupply: capital supplies at firm level and to the stateG: exogenous level of globalization

What is the impact of significant increases in off-shoring, as we may experience

in the next decade or two? We can illustrate a large increase in offshoring in the diagram

by an exogenous shift to the right in the level of openness or globalization. At this higher

level, there is a greater demand for social protection, but also less ability and willingness

of the government to supply it. How will this discrepancy be resolved? It depends on the

bargaining power of the corporations with increasingly more sourcing options, relative to

the bargaining power of the state to raise revenue, spend and regulate, and relative to

38

workers as they make wage and benefit demands. Figure 2, also taken from Braunstein

and Epstein (2001) illustrates the case where bargaining power of the state declines as

openness increases, in this case, as offshoring becomes much more widespread as

predicted by Blinder and others. By this logic, the state is less able and willing to provide

the social protection desired (demanded) by the population as they face higher disruption

and transition costs due to offshoring.

The main point of this discussion is that the bargaining power of these groups is

not independent of the degree of capital mobility our sourcing options. If these

relationships capture some truths of how governance possibilities evolve with

globalization, then the impact of off-shoring needs to be evaluated differently than is

typical in economics discussions. Usually, economists split allocational from

distributional discussions: typically, economists would argue, first and foremost, that off-

shoring will improve efficiency. Then, as a secondary consideration, and if necessary,

winners can compensate losers, or a social safety net can be created to help the displaced,

or firms can be made to buy wage insurance for this purpose. But what if off-shoring

itself makes it difficult or impossible – in a political economy sense—for the state to

undertake these policies? Then what? This, in fact, may be the situation we face. In this

case, what is to be done?

I would put the argument this way: if the policies that economists themselves

propose are to be implemented – that is, social safety nets created, wage insurance

implemented, more or better education financed -- something has to be done to enhance

the bargaining power of the state and workers vis a vis capital as globalization intensifies

and offshoring increases. This is required because the very process of globalization is

reducing states’ and citizens’ bargaining power, arguably below the point at which they

have sufficient leverage to enact these policies that economists propose.

I cannot develop here a whole list of policies that could help. Such a list might

include a wide range of policy and institutional changes including campaign finance

reform, lobbying reform, labor-law reform, to name just a few. But for reasons of space

and also my particular areas of knowledge and interest, I will briefly mention just a few

policy areas to enhance the bargaining power of labor and the state in this arena, and

39

thereby make it more likely that the demand for social protection will be either less

needed, or will be supplied.

Figure 2Effects of globalization on social protection

when it favors capital’s bargaining power

SocialProtection

Openness

G0 G1

Demand

Supply

G2 G3 G4

SP0

ContractCurve

•• • ••

More Expansionary Macroeconomic Policies Can Help

As our statistical analysis on manufacturing outsourcing suggested, aggregate

demand at home and abroad has a large impact on employment, and therefore - by

extension to standard bargaining models -on bargaining power of workers.

But rather than being aggregate demand expanding, macroeconomic and

financial policy in many parts of the globe has been guided by a focus on fiscal austerity,

and fighting inflation, sometimes including formal inflation targeting by central banks. In

the U.S. and Europe, there has been some relaxation of the fiscal constraint in the last

40

several years, but the monetary constraint has remained tight. In the rest of the world,

however, under IMF guidance and the promotion of macroeconomists trained in US and

UK universities, inflation targeting and fiscal austerity has been the norm. (Epstein, 2004;

Pollin, 2005). Restrictive macroeconomic policy in much of the world has made many

countries focus on running export surpluses to generate employment and profits, and on

attracting off-shoring contracts and FDI for the same reason, rather than developing the

home or regional markets as the prime source of demand.18 Compounding the export

orientation and lack of aggregate demand emanating from the developing world has been

the increased mobility and instability of international short term financial flows, along

with the pressure again from the IMF and elsewhere for countries to eliminate their

barriers to capital flows. This has put many so-called “emerging markets” in a defensive

mode, leading them to accumulate large amounts of external reserves to prevent another

financial crisis like the “Asian financial crisis” of the late 1990’s.

This focus on attracting FDI and other export oriented production prospects such

as off-shoring has led to enormous competition for capital and production contracts from

MNC’s in a context of stagnating internal aggregate demand. This only serves to raise the

bargaining power of capital and lower the bargaining power of labor and states.

Alternative macro-economic policy and macroeconomic policy advice that would

promote a greater emphasis on expanding domestic demand, managing capital flows

through capital management techniques (eg., Epstein, Grabel and Jomo, 2005) and

orienting monetary policy toward domestic employment expansion as well as fighting

inflation, could help rebalance power between capital, labor and the state. Such policies

would require a re-thinking of macroeconomic policy advice by the IMF, the U.S.

Treasury and many western trained mainstream economists. (eg. Epstein, 2004; Pollin,

2005).

Domestic and International Tax Treaties

18 China to some extent is an exception to this model where they have pursued export-led growth for other reasons.

41

While there is still debate about the empirical strength and theoretical basis for the

effect, many economists agree that tax competition in the face of mobile capital is likely

to undermine the ability of states to collect tax revenue (eg. Altshuler and Grubert, 2005).

Such competition can occur within federal systems such as the United States, and it can

also occur internationally. It might appear that only international tax competition is

relevant to our discussion, but the so-called “War Between the States” as economists at

the Federal Reserve Bank of Minneapolis call it, that is, the attempt of competing states

to attract investment, can greatly exacerbate international tax competition. (Federal

Reserve Bank of Minneapolis, 1994).19 To counter this trend, some have argued for tax

harmonization policies to reduce the destructive forms of tax competition, both at

domestic and at international levels. (Tanzi, 1995; 1999) Efforts along these lines have

been implemented in Europe and some have argued for their extension to the global

sphere. These policies could significantly re-balance bargaining power in a way that

would be conducive to building the public safety net. There has been some impetus at

international cooperation with respect to money laundering, and in Europe, progress has

been made in terms of restricting subsidy competition to attract investment (Muti, 2003).

Evidently, though, more needs to be done to prevent the erosion of the state’s willingness

ability to compensate losers from offshoring and globalization more generally.

International Rules of the Game in Regulating MNC’s

International rules limiting the ability of national governments to limit the

activities and regulate multinational corporations have been increasingly imbedded in bi-

lateral, regional and international agreements, for example at the World Trade

Organization (WTO) (UNCTAD, World Investment Report, various years). Some of

these agreements have to do with protecting intellectual property rights, limiting

performance requirements on firms, and limiting other forms of regulations. Some have

reduced the ability of countries to use payments to attract investment but most have

interfered with the countries ability to regulate the investment.

19 Currently there is a Supreme Court case testing whether such tax breaks are constitutional in the U.S.

42

On balance, these restrictions are likely to enhance the profitability for MNC’s of

investing or sourcing in developing countries, and thereby, increase their “reservation

price” of staying at home. This outcome would especially be the case when developing

countries increasingly depend on such investment and sourcing for jobs in a world of

austerity and financial instability as I described above. There is evidence that many

developing countries are pressured to accept such restrictions, as part of WTO accession

or borrowing from the World Bank or IMF. Creating more flexibility in these rules, and

allowing developing countries to adopt more restrictions on investment if they choose to

do so, would help to create more balance in bargaining power between labor, capital and

the state in both the richer and the poorer countries.

Threat Effects in Reverse

How can governments and citizens convince companies to accept such policies?

Crotty and Epstein (1997) argued in a different context that capital controls, or the threat

of imposing them, could be helpful, or even necessary to convince firms to pursue more

socially desirable behaviors, such as paying more taxes, increasing domestic investment,

and generating more employment. As long as firms know they always have an exit

strategy, then it is difficult for government or citizens to gain the bargaining power

necessary to win concessions from them.

A similar argument is true in the case of offshoring. Governments and citizens

have to retain the ability to impose or threaten to impose restrictions on off-shoring and

other forms of trade, in order to preserve the bargaining power necessary to be able to

perform the roles of the state that citizens need, including maintaining an adequate social

safety net.

Imposing such trade restrictions would not be considered by economists to be

“efficient” in a static model of perfect competition. But, in a world where allocation,

distribution and power are so inextricably linked as they are in the areas of trade and off-

shoring, the separation of these policies into neat, different piles will not work. Where

trade and trade policies themselves affect the bargaining power of firms, states and

citizens, and these changes affect the feasibility of compensating losers from

globalization, these factors have to be taken ito account in designing policy.

43

Economists have not shied away from using ‘political economy’ arguments when

they want to discuss “rent seeking” and “state failure”. Nor should they shy away from

“political economy” when assessing trade and compensation policy. Economists can

convince themselves that they can talk about the efficiency of off-shoring and then, as an

afterthought, describe a re-distributional package; but if the new political economy has

taught us anything, this artificial separation is wishful thinking, at best, and, more likely,

it contains a heavy dose of self-delusion.

44

Selected Bibliography

Altshuler, Rosanne, and Harry Grubert. 2005. “The Three Parties in the Race to the Bottom: Host Governments, Home Governments and Multinational Companies”. CESIFO Working Paper No. 1613, December. Baily, Martin and Robert Lawrence. 2004. “What Happened to the Great U.S. Job Machine? The Role of Trade and Electronic Offshoring”. Brookings Papers on Economic Activity, 2. Baker, Dean, Gerald Epstein and Robert Pollin, eds.1998a. Globalization and Progressive Economic Policy, Cambridge: Cambridge University Press. Barnet, Richard J. and John Cavanagh. 1994. Global Dreams; Imperial Corporations and the New World Order. New York: Simon and Schuster. Bhagwati, Jagdish. 2004. In Defense of Globalization. Oxford: Oxford University Press. Bivens, Josh. 2004. “Shifting Blame for Manufacturing Job Loss”, Economic Policy Institute Briefing Paper, # 149. Blinder, Alan S. 2005. “Fear of Offshoring”. Princeton University: CEPS Working Paper, No. 119. Braunstein, Elissa and Gerald Epstein. 1999. “Towards a New MAI”, in Jonathan Michie and John Grieve Smith., eds. Global Instability and World Economic Governance. New York: Routledge Press. Bronfenbrenner, Kate, 1997. The Effects of Plant Closing or Threat of Plant Closing on the Right of Workers to Organize, Report to Department of Labor, September, 1996, mimeo. Bronfenbrenner, Kate, 2000. "Uneasy Terrain: The Impact of Capital Mobility on Workers, Wages and Union Organizing". Report to The U.S. Trade Deficit Review Commission. mimeo. Cornell University. Burke, James and Gerald Epstein. 2002. "Threat Effects and the Internationalization of Production". in Jayati Ghosh and C.P. Chandrasekhar, eds. Work and Well-Being in the Age of Finance, New Delhi: Tulika Books. Burke, James and Gerald Epstein. 2006. “Rising Foreign Outsourcing and Employment Losses in U.S. Manufacturing, 1987-2003.” mimeo.

45

Campa, Jose and Linda Goldberg, 1997, “The Evolving External Orientation of Manufacturing Industries: Evidence from Four Countries”, National Bureau of Economic Research, Inc., NBER Working Papers: 5919, February. Chandrasekhar, C. P., and Jayati Ghosh, 2006. “IT-Driven Offshoring: The Exaggerated ‘Development Opportunity’. www.networkideas.org Chang, Ha-Joon. 2004. Kicking Away the Ladder. London: Anthem Press. Choi, Minsik. 2004. “Threat Effect of Foreign Direct Investment on Labor Union Wage Premium” www.umass.edu/peri . Crotty, James and Gerald Epstein. 1998. “In Defense of Capital Controls”. Socialist Review. Crotty, James, Gerald Epstein, and Patricia Kelly 1998. “Multinational Corporations in the Neo-Liberal Regime,” in Dean Baker, Epstein and Pollin, eds. Globalization and Progressive Economic Policy. Cambridge: Cambridge University Press. Epstein, Gerald. 2000. “A Survey of Research on Threat Effects”, University of Massachusetts, PERI. www.umass.edu/peri Epstein, Gerald. 2004. “Alternatives to Inflation Targeting Monetary Policy”, www.umass.edu/peri . Epstein, Gerald, Ilene Grabel and Jomo, K.S. 2005. “Capital Management Techniques in Developing Countries”, in Gerald Epstein, Capital Flight and Capital Controls in Developing Countries. Northampton, MA: E. Elgar Press. Federal Reserve Bank of Minneapolis, 1994. “Congress Should End the Economic War Among the States,” http://woodrow.mpls.frb.fed.us/sylloge/econwar/war-cite.html .Feenstra, Robert C. 1998. “Integration of Trade and Disintegration of Production in the Global Economy”. Journal of Economic Perspectives. Fall, 1998, vol. 12, No. 4, pp. 31-50. Feenstra, Robert C. and Gordon H. Hanson, 1996a. “Globalization, Outsourcing and Wage Inequality,” American Economics Review, May, pp. 240-245. Feenstra, Robert C. and Gordon H. Hanson, 1996b. “Foreign Investment, Outsourcing and Relative Wages,” in Robert C. Feenstra, Gene M. Grossman and Douglas A. Irwin, eds., Political Economy of Trade Policy: Essays in honor of Jagdish Bhagwati. Cambridge, MA: MIT Press, pp. 89-127.

46

Feenstra, Robert C. and Gordon H. Hanson, 1999. "The Impact of Outsourcing and High-Technology Capital on Wages: Estimates for the United States, 1979-1990," Quarterly Journal of Economics, August. Freeman, Richard B. 2005. “Does Globalization of the Scientific/Engineering Workforce Threaten U.S. Economic Leadership?” NBER Working Paper, No. 11457. Freidman, Thomas. 2004. The World is Flat. Greider, William. 1997. One World; Ready or Not. New York: Simon and Schuster. Holmes, Thomas J. 1995, “Analyzing a Proposal to Ban State Tax Breaks to Businesses,” Federal Reserve Bank of Minneapolis Quarterly Review, pp. 29-39. Jensen, J. Bradford and Lori G. Kletzer. 2005. “Tradable Services: Understanding the Scope and Impact of Services Offshoring”, Brookings Trade Forum, 2005. Madrick, Jeff and William Milberg. 2006. “A Proposal for a National Safety Net”. New School University, SCEPA, mimeo. Millberg, William, Rudi von Arnim, Melissa Mahoney and Markus Schneider. 2005. “Spurring Growth Dynamics from Services Offshoring”. Schwartz Center for Economic Policy Analysis. Muti, John. 2003. Foreign Direct Investment and Tax Competition. Washington: Institute for International Economics. Pollin, Robert N. 2005. Contours of Descent. London: Verso Press. Rodrik, Dani. 1997. Has Globalization Gone Too Far? Washington: Institute for International Economics. Rodrik, Dani. 1999. “Globalisation and Labour, or: if Globalisation is a bowl of cherries, why are there so many glum faces around the table?” in, Richard E. Baldwin, et. al., eds., Market Integration, Regionalism and the Global Economy.Cambridge: Cambridge University Press and the Centre for Economic Policy Research. Tanzi, Vito, 1995. Taxation in an Integrating World. Washington: Brookings Institution. Tanzi, Vito. 1999. “Is there a Need for a World Tax Organization” in Razin and Sadka, 1999, pp. 173-186. UNCTAD. 2000. World Investment Report (WIR) 2000. Wood, Adrian. 1995. “How Trade Hurt Unskilled Workers,” Journal of Economic Perspectives. Volume 9, No. 3, Summer, 1995, pp. 57-80.

47